Detecção e classificação de faltas em linhas de transmissão utilizando transformada Wavelet e rede lógica neurofuzzy com aprendizado participativo

Abstract

In the Power Electric Systems the transmission line is the most vulnerable element, since they are subject to faults caused by external and internal factors. Protection systems are employed to minimize the impacts caused by the faults and, currently, digital relays meet thelead role in the fault diagnosis. Several algorithms for fault diagnosis can be used in digital relays and, recently, researches have focused on the use of techniques of signal analysis and intelligent systems, in an attempt to overcome the disadvantages of conventional methods. This work presents a methodology for fault detection and classification of shortcircuit and open circuit faults on transmission lines. The proposed methodology uses the Wavelet Transform for detection and uses the Logic Neurofuzzy Network for classification of the fault, from the extraction of information of current and voltage signals of the transmissionline. In the training of the Logic Neurofuzzy Network was included the Participatory Learning in step generation of membership functions of fuzzy subsets. The algorithms were implemented in a Fault Detection and Classification System and results obtained through simulations had demonstrated the robustness and efficiency of the methodology proposal.